An Analytic Hierarchy Process-based Clinical Decision-making For Dry Eye Disease Diagnostic Test Selection

Purpose: Various diagnostic tests are available to diagnose dry eye disease (DED). It is recommended that the diagnosis of DED be made based on the presence of symptoms and at least one positive homeostasis marker result from the diagnostic test. To date, there is no single gold standard clinical sign that correlates perfectly with DED. This leads to uncertainty for health practitioners in selecting the tests to diagnose DED. The analytic hierarchy process (AHP) is a quantitative decision-making technique for multicriteria decision problems. In this study, a systematic framework using AHP was developed for the multicriteria evaluation of alternatives for DED diagnostic test selection.
Methods: AHP approach was used, where the pairwise comparison of the multiple criteria and alternatives were performed to identify the most preferred DED diagnostic test. After having a comprehensive literature review, four criteria namely: accuracy, reliability, cost, and time needed are important in deciding DED diagnostic tests and seven alternatives (questionnaire, TBUT, NIBUT, Schirmer test, tear osmolarity, Meibomian gland assessment, and ocular surface staining) were used to demonstrate the process. The step began with structuring the decision problem in the form of hierarchy framework where the goal was located at the highest level, criteria in the intermediate level, leaving the decision alternatives at the lowest level. The step was followed by pairwise comparison using Saaty’s nine-point scale to evaluate the relative importance of criteria in selecting DED diagnostic test. Six experts in optometry were recruited to derive the relative importance of criteria and the relative preference of alternatives via questionnaire. The geometric average of experts’ response was obtained. The weight between all criteria and diagnostic tests were calculated.
Results: Accuracy was the most important criteria (weightage=0.62) to be considered in selecting DED diagnostic tests, followed by reliability (weightage=0.23), time needed (weightage=0.09), and cost (weightage=0.06). Schirmer test was the most preferred DED diagnostic test (weightage=0.20) whereas questionnaire was the least preferred (weightage=0.07).
Conclusions: The developed AHP approach can effectively address the uncertainty of practitioners’ clinical decision-making in selecting DED diagnostic tests. It greatly reduces the associated costs, resources, and time involved in obtaining the diagnosis of DED.